It has long been my contention that systemically important financial institutions should not take risks which could jeopardize the ability of the firm to continue as a going concern even if the chances that such risks would result in massive losses are deemed to be remote. If however, such risks must be taken, the one thing that absolutely cannot happen, is for the firm to impede the efforts of those hired to monitor the risks.
It should be noted at the outset that in the case of Eric Ben-Artzi's whisteblower claim, Deutsche Bank contends that Mr. Ben Artzi's dismissal was unrelated to his diligent efforts to critique the firm's handling of a massive synthetic CDO position during the financial crisis. It is my contention however, that Mr. Ben-Artiz's case is evidence of a corporate culture at Deutsche Bank that can result, in some cases, in the facilitation of extraordinarily risky behavior and in the stifling of attempts to monitor the risks of that behavior by discouraging transparency. This type of corporate culture is not one in which investors should place their faith by way of share ownership. This article is meant to provide support for my contention.
Thanks to Eric Ben-Artzi and his whisteblower claim against Deutsche Bank (DB), the concept of an LSS 'gap option risk' (previously obscure outside of the credit derivatives world) has become an unlikely buzzword around the water cooler. It was this gap risk - the theoretical losses which would occur in the event a credit protection writer walked away from a leveraged super senior (LSS) swap deal leaving behind collateral worth only a fraction of the notional amount, effectively leaving the buyer partially naked vis-à-vis similar offsetting swaps - that Ben-Artzi claims was improperly modeled at Deutsche, allowing the bank to cover up over $10 billion in paper losses during the financial crisis.
While I wouldn't go so far as to say this concept is now well understood, I would say that Mr. Ben-Artzi's case has resulted in heightened public awareness of the issue and that is probably a good thing. In addition to the fact that financial institutions will most likely be more vigilant in their gap option risk modeling as a result of the claim, investors' collective level of sophistication has been made greater by learning about the trade and an informed investing public makes better decisions than an uninformed investing public.
There is however, another important contextual issue germane to this case that hasn't received as much attention but once understood, further demonstrates just how risky the situation actually became. The Deutsche Bank trades were part of a group of similar trades involving foreign financial institutions and Canadian special purpose vehicles and there was a point, in the autumn of 2007, when it appeared that many of these LSS trades might result in complete losses - the swaps cancelled, the collateral gone.
This back-story was largely (and conspicuously) absent from discussions and explanations of the Deutsche Bank trade until Yves Smith of Naked Capitalism accused DealBreaker's Matt Levine and Reuters' Felix Salmon of "airbrush[ing] out what has been called the biggest credit restructuring in history." For at least one well-known financial institution, this restructuring did not end happily.
The implicit argument made my Deutsche's defenders is that once the trades are put in the proper context - which, to them means considering that it was 'crisis time' and the trades were only "paper losses" - one can understand why what the bank did was the right thing under the circumstances. In reality, the background story behind these trades demonstrates that they were perhaps far more risky than even Deutsche's accusers allege. For instance, in my own correspondence with Mr. Ben-Artzi, he indicated that,
"...in modeling and valuing the trades internally, it wasn't clear if the credit risk adjustment for the collateral had to be accounted for. It was clear that the value of the unwind option, i.e. the option to walk away rather than post additional collateral, had to be valued."
The credit risk associated with the collateral hasn't received as much attention as the gap risk. However, as the following story will make abundantly clear, Deutsche Bank and other so-called "asset providers" faced not only the gap option risk associated with the early termination of the swaps, but, thanks to an unprecedented disruption in the market for Canadian asset backed commercial paper, they also faced the loss of the collateral posted to secure the trades and the prospect of making liquidity payments to cover their counterparties' default on commercial backed paper.
Barclays Gets A Notice
It's August 15, 2007 and Barclays has, for the third consecutive day, received a "disruption notice" from Devonshire Trust in Canada. To be sure, neither Barclays nor Devonshire ever thought it would come to this but ultimately, the unthinkable had happened and Devonshire needed money to pay investors who had bought commercial paper from the Trust - paper which had now come due. Normally, the investors would be paid with money from newly issued paper, or, they would simply roll their debt forward. Not this time. Jitters about the possibility that some of the "assets" behind the asset backed commercial paper (ABCP) could be tied to subprime lending had effectively caused the Canadian ABCP market to freeze on August 13 and now, Devonshire expected Barclay's to honor its commitment to provide liquidity to investors in the event of a market disruption.
Turn back the clock nearly two years to better times. In October 2005, two former Deutsche Bank employees, Mathieu Lafleur-Ayotte and Alain Pelchat, had just left their latest positions as managing directors of National Bank's structured finance division to start "Quanto," a so-called conduit business through which, with the help of another former Deutsche employee John Lovisolo, the two negotiated a pair of transactions with Barclays. The transactions were actually conducted between Barclays and "Devonshire Trust" a special purpose vehicle for which Quanto acted as the financial services agent.
Devonshire's special purpose was, in the words of Ontario Superior Court Justice Frank Newbound, to "acquire and hold income-producing assets financed through the issuance of asset backed commercial paper." Here, in plain English, is how this worked. Devonshire agreed to sell Barclays protection on a portfolio of debt obligations. That protection was effectively collateralized, or partially guaranteed, by the money Devonshire received from the investors who bought the commercial paper. In return for this protection, Barclays paid a monthly fee (like an insurance premium) to Devonshire, which was used to make interest payments on the notes the Trust issued (the commercial backed paper) to finance the trade. Put more succinctly: Devonshire sold commercial paper and used the investors' money to guarantee the protection it sold to Barclays, and the fee Barclays paid for that protection was used to make what amounted to interest payments to the buyers of the commercial paper.
A Synthetic, Leveraged Super Senior Deal
Keep in mind that Barclays did not own the underlying debt obligations - it was buying protection against defaults on something it did not own. Why? Because, as noted by the Ontario Superior Court, Barclays was a swap dealer and the idea here, presumably, was to make money from the spread between the premiums it paid for the protection it bought from Devonshire and the premiums it collected from the protection it sold to other investors, while keeping itself (relatively) neutral in terms of payouts. This lack of ownership stake in the underlying is why the deal is deemed "synthetic."
There were two swaps involved in this trade and each carried a notional amount of $3 billion which was the amount Devonshire could theoretically be on the hook for in the event some catastrophic credit event occurred. I say "catastrophic" because Devonshire was only liable for payments on losses in the "super senior tranche." Effectively, the underlying bond portfolio would had to have suffered aggregate losses of between 15 and 16% before Devonshire was on the hook - this threshold is known as the "attachment point." Neither party was particularly worried about this scenario. From Justice Newbound:
"Because of the quality of the corporate bonds which made up the underlying portfolio on which Barclays bought credit protection, it was thought by both Barclays and Devonshire that it would be highly unlikely that any losses would occur during the term of the swaps that would reach the attachment point, and thus highly unlikely that Devonshire would ever be called upon to pay anything to Barclays for losses in the super senior tranche."
Probably because of these remote odds, Devonshire was only required to post $300 million in collateral for each swap for a total of $600 million on a notional amount of $6 billion, meaning the trades were leveraged 10-1. To summarize: the trades were predicated upon the super senior tranche of a bond portfolio, the participants didn't own the underlying debt obligations, and the trades were leveraged. This is where the name "synthetic leveraged super senior", or "LSS" swap comes from.
Gap Option Risk
The structure of the deal worked out well for Devonshire because the Trust collected premium payments on $6 billion in sold protection while only posting collateral of $600 million. As of 2016, Devonshire was set to get its collateral back as long as there were no defaults before then and in order to make everyone feel comfortable, Barclays posted $600 million of its assets in a custodial account at the Bank of New York Mellon. So let's recap (and this should provide a bit of comic relief): Barclays posted collateral to secure the repayment of Devonshire's collateral, which itself was derived from commercial paper sold to investors, to whom payments were made out of money Barclays paid to Devonshire in exchange for protection bought on debt obligations which none of the involved parties actually owned.
Now, there were conditions in the contracts which stipulated if and when additional collateral would need to be posted by Devonshire (so-called "triggering events") and here you can see where the 'gap risk' (the main point of contention in the Deutsche Bank case) comes in:
"to protect Barclays in the event that the value of the swaps to Barclays increased beyond certain trigger points based on a mark to market valuation of the underlying portfolio for which credit protection was being purchased, Devonshire agreed to post further collateral to Barclays on certain conditions. This was to protect Barclays against its "gap risk", being the risk that the collateral held by Barclays was less than the loss to Barclays in the event that there was early termination of the swaps." (emphasis mine)
So, as I noted in a previous article, the issue for the protection buyer (in this case Barclays, in the other case Deutsche) is what happens if the protection seller (the Canadians) walks away from the swaps just as credit risk flares up, leaving behind only the originally posted collateral worth just a fraction (1/10) of the total amount insured? This is bad news if the party who in this case is the protection buyer is, in another case, a protection seller of an equivalent or nearly equivalent notional amount. That is, buying the protection was meant to kill two birds with one stone: it protects you in the event that you become liable for losses in a similar portfolio on which you wrote protection (you simply match the payouts you receive with those you must make), and it allows you to profit in the mean time as long as the premium payments you receive as a protection seller exceed those you pay out as a protection buyer.
If however, the counterparty from whom you bought protection were to simply walk away when the going gets tough in terms of the credit markets, you would be forced to either walk away from your own obligations as a protection seller (this is how counterparty risk can spiral out of control) or else take a loss on the difference between the payout to those who bought protection from you and the collateral you were able to keep from the protection seller who left you high and dry. That difference is the "gap risk."
Collateral Risk And A Liquidity Provision
Once the Canadian ABCP market froze, Devonshire found it could no longer roll short-term (30-90 day) commercial paper and was similarly unable to sell additional notes. Because of this, it could not make payments to investors whose notes had come due. The two swap deals with Barclays contained a provision which specified that in the event of a "market disruption" in the ABCP market, Barclays would provide up to $205 million in payments to holders of maturing commercial paper. Add in the fact that Barclays collected a monthly fee from Devonshire for this liquidity guarantee and this circle of financial lunacy is complete: essentially, both parties were on the hook to each other for this or that at nearly every turn and the whole thing was held together by the financial market equivalent of nuclear deterrence theory. Ultimately, when Devonshire asked Barclays to make good on the liquidity provision, Barclays simply denied that the problem in the ABCP market constituted a "market disruption" as defined in the swap agreements and on those grounds, refused payment.
"More than 20 Canadian ABCP trusts were frozen in August 2007 when they were unable to roll over maturing paper during the turmoil in the US subprime mortgage market. A group of Canadian and international banks that acted as swap counterparties to the trusts refused to provide extra liquidity on the grounds that the ABCP market had not been disrupted by the financial meltdown.
As credit spreads blew wider, the banks were set to make collateral calls on the Canadian trusts and the Canadian trusts couldn't roll any more paper so they were set to make liquidity calls on the banks. True to the above-mentioned nuclear analogy, everybody was ready to fire their warheads at once (this is reminiscent of the language used in the court case) setting off a cascade of disputed liquidity provision payments on one side and a wave of additional collateral postings on the other side.
In order to avoid this nightmare scenario, everyone (well, everyone except the trusts) met in Montreal on August 16, 2007 to sort the whole thing out. Out of this meeting the Montreal Accord was born and among its first provisions was a 60-day moratorium on calls for liquidity payments and on demands for the posting of additional collateral - a cease-fire if you will.At this time it was understood that the Canadian trusts would need to be restructured and, to make a long story short, Barclays did not agree with the restructuring terms developed and discussed within the Montreal Accord and hostilities between Devonshire and Barclays commenced.
The culmination of the bad blood came on January 13, 2009 when Barclays sent notice to Devonshire of the termination of the swap agreements on grounds that Devonshire was insolvent and simultaneously wired Devonshire the $71 million it owed the trust for the liquidity payments Devonshire originally requested a year and a half earlier. Devonshire refused to accept Barclays' termination of the swap deals and later that same day, sent Barclays a notice of termination.
It turned out that Devonshire's case made the most sense: Devonshire contended that Barclays had no right to terminate the swap deals on the basis of Devonshire's insolvency because it was, after all, Barclays who caused the insolvency in the first place by failing to make liquidity payments. Thus, termination of the swap deal by Barclay's amounted to profiting from an event it had itself caused. In the end, Barclays lost the collateral initially posted by Devonshire. From the case:
"…an event of default under section 5(a)(I) occurred at the close of business on January 12, 2009. I have also held that Barclays' payment on January 13, 2009 did not cure its default. Thus Barclays for the purposes of this trial was a defaulting party under section 6(a) and Devonshire had the right to deliver its early termination notice designating January 13, 2009 as the termination date…Pursuant to the Intercreditor Agreement, Devonshire is entitled to be paid the $600 million delivered by it to Barclays at the outset of the transaction." (emphasis mine)
Devonshire was the only conduit not successfully restructured during or after the Canadian ABCP crisis. The point however, and I'll quote Yves Smith again here, is that it very well could have been Deutsche (or any other protection buyer involved in these deals) taking losses right alongside Barclays:
"If the restructuring had failed, Deutsche most assuredly would have had major losses on its hands, not just from being forced to unwind the trades at an unfavorable time, but also from litigation, which is not factored into the whistleblowers' estimates..And the restructuring failing was not a hypothetical risk. There was not assurance it would get done, and analysts with no dog in the fight were of the view at the time that it was not likely to be completed successfully."
Ultimately though, the restructuring was completed and C32 billion in commercial paper was converted to long-term, floating rate notes and packaged into three so-called "Master Asset Vehicles" (MAVs). MAV I and MAV II held some 75% of their assets in leveraged CDOs. These vehicles are still around today and MAV I and MAV II were in fact upgraded by Canadian ratings firm DBRS (the same DBRS who kept Spanish T-Bills at "low A" in October allowing the ECB to claim that it didn't improperly haircut Spanish bonds posted as collateral in liquidity ops) on June 27 and June 28 respectively.
Bringing It All Together
One can see, from Barclays' experience, what very well could have happened if the Montreal Accord had not succeeded. Disputed liquidity payments might well have led to extended legal battles that, in the end, would have very likely resulted in the termination of the swaps (presumably leaving "asset providers" like Deutsche unhedged) and could have ended with a total loss of the posted collateral as well, leaving nothing but naked short positions and legal fees for Deutsche and the rest.
Additionally, it is clear that the conduits could demand the liquidity payments from the banks. What isn't clear is whether the banks could realistically demand the offsetting collateral calls from the conduits. Presumably a market disruption sufficient to trigger the liquidity provisions would also trigger collateral calls, but since the failure of Barclays to pay out the commercial paper holders essentially bankrupted Devonshire, one wonders if the idea that these special purpose vehicles could have actually come up with the collateral to post was not ridiculous on its face.
It should also be noted that this issue is not, as some would have you believe, dead and gone. Although I certainly do not make any claim to knowing the details of the deals that are currently in place, what can be surmised from Deutsche's 2011 annual report is that the bank is still involved in these deals - the bank is a liquidity provider to asset backed commercial paper issuers:
"As an administrative agent for commercial paper programs, we facilitate the purchase of non-Deutsche Bank Group loans, securities and other receivables by the commercial paper conduit (conduit), which then issues to the market high-grade, short-term commercial paper, collateralized by the underlying assets, to fund the purchase. The conduits require…liquidity support to maintain an investment grade rating for the commercial paper. We are the liquidity provider to these conduits and therefore exposed to changes in the carrying value of their assets...Our liquidity exposure to these conduits is to the entire commercial paper issued of € 11.6 billion and € 16.3 billion as of December 31, 2011 and December 31, 2010, of which we held € 2.5 billion and € 2.2 billion, respectively." (emphasis mine)
Finally (and this perhaps the most important point of all), investors should understand that the notional value of Barclays' two swap agreements was a combined $6 billion. The notional amount of Deutsche's swaps was more than twenty times larger at $130 billion. According to Ben-Artzi, valuing the gap option properly would have meant losses totaling $10.4 billion or, around a fourth of Deutsche's capital at the end of 2008. However, had a fate similar to Barclays' befallen Deutsche Bank (i.e. in the course of litigation the swaps were cancelled and the collateral returned to the conduit) the combined losses between the realization of the gap option risk and the returned collateral could well have totaled some $20 billion, nearly half of the Deutsche's capital at the time.
This is not some wildly unlikely theoretical eventuality derived from a tortured examination of ancient history. When the Canadian asset backed commercial paper market froze in 2007, these LSS trades were at a very real risk of imploding. As the Barclays case shows, asset providers (protection buyers like Deutsche) were perilously close (one failed summit in Montreal) to litigation which could have, in all cases cost them dearly, and in Deutsche's case, could have brought the firm to its knees.
As noted at the outset, investors should understand the risk-taking propensity of the firms in which they have ownership. Deutsche Bank claims it has done nothing wrong in this case. I believe investors should ask themselves what this says about the bank's willingness to take similar risks in the future. Furthermore, as indicated by the above cited passage from the firm's 2011 annual report, the bank is still doing these deals. It would seem nothing has been learned from what, in my judgment, is an extraordinarily close brush with collapse. In fact, the firm hasn't even acknowledged any wrongdoing whatsoever. In my opinion, this case proves, beyond a shadow of a doubt, that Deutsche Bank's shares are not where a conservative investor should have his money tied up.
Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.